* cache loads across buffers (since they may share rawbufs)
* typing
* add test
* fix test
* small changes to test
* fix test
* one big cache
* whitespace
* golf a line?
* invalid is RawBuffer(0)[0], valid 1.
* Symbolic Shape JIT
update tests
2 variables symbolic ops, adding more tests
test passing
cleanup
* more test cases
* single flag
* review update
* jit attention one piece
* realize
* symbolic_jit test for cuda
* old artifact
* works with cuda gpu but failed ci
* CUDACPU
* move assembly, assembly_ptx
* successful but broken rendering of ptx asm
* clear ins before render asm
* slightly less broken :')
* we needed thread syncs
* fix float16 loading, rounding modifiers and other casting stuff, passing casts_from_half
* Fix runtime_args for gpuocelot
* our casts were flipped on both ends
* more casting
* add ternary where op
* dealing with storing/loading bool
* add test for casting to bool from negative
* Fix args.valid on ConstOp
* add to CI, TODO: fix runtime_args for test_uops
* fix placement of runtime_args to work with lazy.Device
* undo ci changes so I can push
* fix lints
* start cleanup and fix things we broke fixing lints
* add checks for PTX specifc asm instructions
* revert added test -- doesn't pass on llvm
* skip tests for underflow,overflow
* another fix for how we're setting runtime args
* Less broken cleanup
* add to CI
* add more env variables for ci test
* fix ci to install pycuda for ptx
* ci: copy cuda test command
* cleanup
* assert to make sure we're actually running ptx in ci
* remove test assert
* move is_ptx arg
* move assembly, assembly_ptx back to extras
* fix imports
* initial merge fixes
* clear registers, fix UOps.LOAD with invalid value
* draft merge fixes
* remove prints
* quick lint and merge fixes
* cleanup
* remove PTXProgram wrapper
* final cleanup
* temp change for ci rerun
* ci rerun
* rollback ISA version
* try to run commavq
* fix 0 dim, start implementing new ops
- Implement EmbedLayerNormalization
- Implement Attention
* SkipLayerNormalization and FastGelu
* use original torch model, cast inputs
* fix some ops:
- properly do Cast
- Attention: bi- and unidirectional
- FastGelu: add bias before gelu
* cleanup onnx_ops.py
* add validation option to benchmark
* cleanup imports
* add checks incase onnx2torch implements ops in future
* run onnx instead of original torch
* just skip gpu on m1
* reactivate the other models
* check for strange params & squash whitespace
* cleanup
* fix causal mask Attention
* Range doesn't need int cast
* embedding vocab_counter same dtype as input
* no need to cast
* always validate, fix PosixPath ort
---------
Co-authored-by: George Hotz <george@comma.ai>
* testing new memops
* better debugging
* testing padded conv
* branching with load
* refactoring a bit
* first try
* fixing bugs
* fixing some
* eq
* eq2
* do not use x's
* working
* fixing imm
* getting things working
* refactor
* pow not working
* working except one
* refactor: one store mem
* refactor: global load
* refactor: imm
* refactor: cleaning
* fixing big offsets
* refactor with ci
* try ci
* typo
* another typo
* ubuntu default
* forgot git
* do i need git?
* missing packages
* adding python-dev
* with cache?
* buildx action
* buildx name issue?
* maybe now?
* python3
* newline warning
* maybe now
* i actually need this
* ci should work now
* improved caching
* fixing cache
* maybe now it will cache
* this
* testing cache
* trying again
* load
* missing platform
* caching gha
* testing cache
* full testing
* typo
* now?
* why
* adding checkout back
* bad formatting
* fixing convention issues
* supporting python
* adding CI flag
* testing all
* better comments
* adding debugging
* takes 12x longer
* does it output progress now?
* ignore models for speed
* fixing merge
* excluding conv_transpose2d
* only 2 test cuz is to slow
* another approach
* let's see
* faster duh
* my bad
* T_T
* typo
* sup
* with output?
* comment test
* comment test
* comment test
* :?
* no comment
* with cache
* back to normal
* testing that ci works
* back to passing
* trying again
* does it create another entry
* does it create another entry?
* build local
* hey
* Revert "excluding conv_transpose2d"
This reverts commit cc7348de03033e032f47d69caff174e2f1a7bfea.
* does it cache if done before?
* does it cache?
* done
* adding test ops
* bad formatting
* no need for this
* working static mem
* sum 1d
* add ndim
* better reg import
* fix stack
* back to np
* working except for softmax
* 5 failing
* no pogress
* remove keystone
* remove keystone
* testops passing
* cleanups
* more cleanup
* typo
* ci
* ci2
* cond import
* ci3
* ci4
* ci4
* ci5
* ci5
* ci6
* aligment
* test all
* correct test
* err read_unmapped
* passing test
* ignore for speed
* ignore for speed
* ci7
* cleanup
* remove docker
* fixing merge
* fixing bugs
* add skipload for const ops
* comments
* First merge to master: Renderer
* fix emulation
* passing all tests arm64
* cleaning
* fix handcoded binary
* cleaning
* fix errs
* fix runtime arg binary
* clean git diff
* fix and clean
* fixing metal test
* cleaning
* fix metal test
* ci ~8 min
* fix pylint and clang
* cache the files in ops_clang
---------
Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
* do reshaping without merge_views and reshape masks
* added tests
* properly do reshaping of zero or negative masks
* replace while loop with single expression
* remove old condition
* add more tests and comments
* remove empty file
* use scaled attn from Tensor
* add a test for bert
* linter
* no more tokenizer
* without loading weights
* remove prints
* tribute to linter lords
* smaller input and less runs
* small bert
* feat: world
* feat: tests
* feat: no more backwards
* feat: recv into
* feat: whoops
* feat: test in ci
* feat: some debug logging
* feat: workflow naming
* feat: need to set pythonpath
* feat: just send to same device
* feat: allreduce
* feat: test
* feat: need contiguous
* feat: test in ci
* feat: exit with correct code
* feat: don't need that
* feat: opencl wait_for just doesn't work
* feat: synchronize on out
* feat: try?
* feat: try again?
* feat: add extra realizes
* feat: print
* feat: seed
* feat: tol
* feat: test ones and zeros
* feat: remove print
* feat: are you just flaky
* feat: seperate scatter and gather?
* feat: just try synchronizing
* feat: remove print again
* feat: bring back difference
* feat: no sync
* feat: revert that
* feat: back to wait_for
* fix: typo
* feat: world
* feat: tests
* feat: no more backwards
* feat: recv into
* feat: whoops
* feat: test in ci
* feat: some debug logging
* feat: workflow naming
* feat: need to set pythonpath
* feat: just send to same device
* Implement scaled_dot_product_attention and test
* Support attn_mask
* Support is_causal too
* Use in llama
* Don't forget to reshape
* Set requires_grad=False for causal
* Remove staticmethod
* Remove extra spaces
* add disk_tensor
* fix jit
* new baseline before whitening
* whitening through torch
* whiting done currently at 91.65%
* 91.99%
* clean up mixup and 92.3%
* clean up 92.30%
* 92.49% before searching for new hyper-parameters
* fix CI
* fix white space
* add whitening init in test
* refactor, update hyperpara, 92.72%
* converting whiting to tinygrad operation
* update CI kernels count for CIFAR
* add pad reflect
* add random crop 92.53%
* update hyperpara 93%
* 93.15% on docker container, need to refactor the assignment for hyper param
* print out weights and bias to be separated
* bias/non-bias params separated
* fix whitespace
* clean up
* refactor hyper-param with dict
* refactor lr schedular params
* fix whitespace
* fix cross entropy loss
* fix whitespace
* move opt hyp to hyp dict
* minor fixup
* adjust model, loss scaling
* 92.74% while using half of compute as before
* update hyp for cutmix
* random shuffle during batches
* clean up
* updating the model
* update ConvGroup
* disable gradients for batchnorm layer weights
* whitespace
* 93.92%
* clean up
* finally 94%git add .!
* rewrite whitening to remove dependency on torch
* whitespace
* remove dependency on torch, 93.91%
* back to 94.03%
* clean up
* update test_real_world
* feat: add ops_shm
* clean: extra newline
* feat: add test
* feat: ci doesn't like that
* feat: ci still doesn't like that
* feat: skip big test on ci
* feat: testing
* feat: big
* feat: testing again
* feat: reskip test
* Context and Timing can now be used as decorators
* Using Timing decorator in quickstart.md
The time formating is better and is a useful tool to learn.
Old: Time: 3.5260659999912605
New: Time: 3526.14 ms
* Updated env_vars documentation for Context
* Added test for Context decorator
* Put new import on same line as others
* flake8: Ignore frequent violations, correct infrequent ones
* Ignore some rules in test
* Reorder test ignores
* Lint test + main
* EOF indent
* Include all E71,E72 errors
* Test the failing case in CI
* Revert "Test the failing case in CI"
This reverts commit 110add0a70f5a619d07631269104e84f908af6b9.
* Push to test!
This reverts commit f317532779a0e1ac8401e2474fd5c6c8695c08e9.
* ok back to passing
This reverts commit ba5052685f93f83e06152cdc696b9e26131d8ab7.
* Prove that CI fails when formatting is incorrect.
* Fix formatting
* Remove duplicitous E117 rule
* Use flake8 config for precommit
---------
Co-authored-by: waifairer <waifairer@gmail.com>
* Fix max nan
* Adds nan check option to max function
* Calls to max can pass in "ignore_nan=True" argument
* Added max nan CI tests
* Fix max nan
* Adds nan check option to max function
* Calls to max can pass in "ignore_nan=True" argument
* Added max nan CI tests
* Turned off due to the need for granularity
* models matrix
* fix typo and install gpu deps
* install llvm deps if needed
* fix
* testops with cuda
* remove pip cache since not work
* cuda env
* install cuda deps
* maybe it will work now
* i can't read
* all tests in matrix
* trim down more
* opencl stuff in matrix
* opencl pip cache
* test split
* change cuda test exclusion
* test
* fix cuda maybe
* add models
* add more n=auto
* third thing
* fix bug
* cache pip more
* change name
* update tests
* try again cause why not
* balance
* try again...
* try apt cache for cuda
* try on gpu:
* try cuda again
* update packages step
* replace libz-dev with zlib1g-dev
* only cache cuda
* why error
* fix gpuocelot bug
* apt cache err
* apt cache to slow?
* opt and image in single runner
* add a couple n=autos
* remove test matrix
* try cuda apt cache again
* libz-dev -> zlib1g-dev
* remove -s since not supported by xdist
* the cache takes too long and doesn't work
* combine webgpu and metal tests
* combine imagenet to c and cpu tests
* torch tests with linters
* torch back by itself
* small windows clang test with torch tests
* fix a goofy windows bug
* im dumb
* bro
* clang with linters
* fix pylint error
* linter not work on windows
* try with clang again
* clang and imagenet?
* install deps
* fix
* fix quote
* clang by itself (windows too slow)
* env vars for imagenet
* cache pip for metal and webgpu tests
* try torch with metal and webgpu
* doesn't work, too long
* remove -v
* try -n=logical
* don't use logical
* revert accidental thing
* remove some prints unless CI
* fix print unless CI
* ignore speed tests for slow tests
* clang windows in matrix (ubuntu being tested in imagenet->c test)
* try manual pip cache
* fix windows pip cache path
* all manual pip cache
* fix pip cache dir for macos
* print_ci function in helpers
* CI as variable, no print_ci
* missed one
* cuda tests with docker image
* remove setup-python action for cuda
* python->python3?
* remove -s -v
* try fix pip cache
* maybe fix
* try to fix pip cache
* is this the path?
* maybe cache pip
* try again
* create wheels dir
* ?
* cuda pip deps in dockerfile
* disable pip cache for clang
* image from ghcr instead of docker hub
* why is clang like this
* fast deps
* try use different caches
* remove the fast thing
* try with lighter image
* remove setup python for cuda
* small docker and cuda fast deps
* ignore a few more tests
* cool docker thing (maybe)
* oops
* quotes
* fix docker command
* fix bug
* ignore train efficientnet test
* remove dockerfile (docker stuff takes too long)
* remove docker stuff and normal cuda
* oops
* ignore the tests for cuda
* does this work
* ignore test_train on slow backends
* add space
* llvm ignore same tests as cuda
* nvm
* ignore lr scheduler tests
* get some stats
* fix ignore bug
* remove extra '
* remove and
* ignore test for llvm
* change ignored tests and durationon all backends
* fix
* and -> or
* ignore some more cuda tests
* finally?
* does this fix it
* remove durations=0
* add some more tests to llvm
* make last pytest more readable
* fix
* don't train efficientnet on cpu
* try w/out pip cache
* pip cache seems to be generally better
* pytest file markers
* try apt fast for cuda
* use quick install for apt-fast
* apt-fast not worth
* apt-get to apt
* fix typo
* suppress warnings
* register markers
* disable debug on fuzz tests
* change marker names
* apt update and apt install in one command
* update marker names in test.yml
* webgpu pytest marker
* Add additional kernel when reducing multiple dimensions at once.
* Faster for smaller inputs
* Whitespace and naming
* Cleaner, guard for Metal only, and max 1 split rather than N
* Draft of different approach
* One additional kernel call for this test (as expected)
* Fuzz test symbolic and shapetracker
This reverts commit d5773ddebff54c1ff608838076f0b4ff126b8aa8.
* mess again
* no tail
* test shapetracker too
* Revert mess and enable all tests
* removed leftover
* new version
* fix abstractions
* try remove test
* Revert "try remove test"
This reverts commit 2fc18a9f8ed180540baf73d32b568262709822f1.
* assert_allclose
* minimize the test
* minimize the test
* minimize the test
* minimize the test
* Revert "minimize the test"
This reverts commit e0c092959636109f745d1c8a73f2db90c75fe3c1.
* Revert "minimize the test"
This reverts commit 88240551b13403b21a81765043d5736103a49293.
* Revert "minimize the test"
This reverts commit 78328a7ce27328c8bf9a325ae017cc2a4d98f65b.
* Revert "minimize the test"
This reverts commit 989523fded4319b13db047e45ad8c35c861a36aa.
* skip test inside body
* oops
* oops
* Rename FusedOps to TernaryOps
* Support ternary broadcast
* Add where llop and mlop
* Make where op work in cstyle codegen
* Don't skip test_inf_where
* Add backward path to where op
* Use bool in cstyle codegen
* Add LLVM where op
* Add numpy where op
* Add torch where op
* Simplify where mlop
* Update documentation
* Forgot a rename
* Merged relevant changes from PR #1195 onto PR #1196
* Add test to cover changes to linearizer.ast_parse for WHERE op
Without this METAL will try to use ternary op on float4 and fail
* Make where op work in wgsl backend
* Allow ternary ops to be merged
* Make mypy happy
---------
Co-authored-by: Francis Lam <flam@alum.mit.edu>
* WIP: `tensor.squeeze` function
* Added `test_except` param to `helper_test_op` to avoid false positives
* Extracted new method `helper_test_exception` for testing exceptions
* Made `squeeze` not throw IndexError when ndim == 0 and dim <= 0 to match PyTorch
* initial commit
* 81 passing
* 105 passing tests
* 148 passing
* CI tests
* install dep on ci
* try opencl pkgs
* try using vulkan
* down to only 6 failing
* refactor
* cleaning up
* another test skipped due to buffer limit
* linter
* segfault
* indent fix
* another segfault found
* small touchups
* Fix max and maxpool tests
* Add constant folding
* Add javascript export script
* better asserts in codegen
* manual upcasting
* reverted token type change
* skip safetensor test due to unsupported type
* FIx efficientnet and all other model tests
* Remove np copy
* fixed indent and missing import
* manually destroy the buffer
* revert back to length
* linter errors
* removed extra val
* skip broken tests
* skipping more tests
* Make the page pretty
* Save model weights as safetensor
* Fix imagenet to c test
* Fix second imagenet to c bug
* Async and paralel kernel compilation
* workgroup support
* reversed local size
* fixed non local bug
* correct local groups
* ci experiment
* removed typo
* Fix define local by using shared memory
* Refactor
* try running on mac
* match metal tests
* add more workers
* scope down tests
* trying windows runner
* fixed windows env
* see how many it can do
* merged master
* refactor
* missed refactor
* increase test suite coverage
* missing import
* whitespace in test_efficientnet.py
* getting there
* fixed reset
* fixed bufs
* switched to cstyle
* cleanup
* min/max rename
* one more linter issue
* fixed demo
* linter
* testing ci chrome
* add unsafe webgpu arg
* add build step
* remove WEBGPU from cmd line
* use module
* try forcing directx
* trying forced metal backend
* temp disable conv2d for CI
* disable conv_trasnpose2d
---------
Co-authored-by: 0x4d - Martin Loretz <20306567+martinloretzzz@users.noreply.github.com>
Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
* Added test coverage for int32 in `test/test_dtype.py`
Tests for int32 include:
- testing that int32 can be converted into a numpy array
- testing that float and int64 can be cast into int32
- testing that int32 can be cast into float and int64
- testing addition, multiplication, and matrix multiplication with int32
- testing that addition, multiplication, and matrix multiplication with int32 and either float or int64 gets successfully cast into float and int64, respectively
Additional changes include testing that int8 casts into int32 and testing that float16 casts into int32
* Added type casting to the add, subtract, and divide binary operations
* Added automatic type casting when types differ to FusedOps.MULACC
I moved the match_types function back so that I could call it in einsum_mulacc where it would cast the types of the MULACC to be the same
* Added unit test for match_types and added type hints to the parameters
* Added tests for ops_cpu.match_types
* Changed ops_cpu.einsum logic to play nicely with PyTorch
Changed `tinygrad.runtime.ops_cpu.einsum_mulacc` logic to not perform type matching. Type matching was instead moved to the numpy_fxn_for_op dictionary in the ops_cpu file. Since ops_torch uses the same einsum_mulacc function, this should fix all the broken pytorch tests.
* empty commit to rerun ci
* reverting PR#1213 in attempt to fix broken test
* Removed all tests I added to see if they are causing CI issues
* Added back type matching tests
* removed type matching tests and added back int tests
* added back part of the type matching tests
* removed braking type matching tests
* empty commit for testing
* added test back but inside comment
* removed a test from the comment to see if it breaks CI
* removed another function
* more testing
* emptied test comment
* cleaned up comments
* Added optimize=True flag to einsum_mullac in cpu_ops.py
* Removed unnecessary imports from tests
* optimized match_types by removing unnecessary array copying
* Rename in files
* Move files
* Moved to extra/datasets as suggested
* Changes to files
* Fixed stupid mistake
---------
Co-authored-by: terafo <terafo@protonmail.com>
* Fixes + improved test coverage for helpers.py
- added exception handling in `proc`, if an exception was thrown, the thread would hang
- made `_early_exec_process` catch any Exception, before if an exception was thrown before the process was started, it would hand the thread
* Made `_early_exec_process` catch any Exception
Otherwise, if an exception was thrown before the process was started, it would hang the thread. For example a type error for an argument passed to `subprocess.check_output`
* Fixed `from tinygrad.helpers import Timing` import
oops, for some reason my IDE cleaned that import from extra/helpers.
* Fixed import in llama.py
Another one that I skipped by accident, mybad
* Extracted a class for tests of early exec
* Normalize line endings, windows uses /r/n
* Made `cross_process` not a daemon
* fold expands that precede a reduce if the reduction is on the same axis as the expansion
* add deterministic test for SIMPLIFY_SUM_RESHAPE_EXPAND_SUM optimization
* add a test case to make sure we don't fold reduce-expand-reduce on different axes
* fixed division by zero for fast operations
* made et closer to 0
* replace POW llop with SQRT
* updated mlops to swap SQRT and POW llops
* updated hlops to swap POW and SQRT
* added sqrt llop to cpu runtime
* added sqrt llop to cstyle codegen
* added POW llop to llvm ir codegen
* added SQRT llop to torch runtime
* moved pow from mlops to hlops
* found a better way to do reverse pow
* fixed indentation
* added SQRT llop to triton
* update docs to match new llops
* removed POW operator from assembly codegen
* added sqrt and rsqrt to pow hlop
* rewrote pow function in tensor.py
* Adjust tolerance
* Adjust for adamw
* Reduce for Adam too
* removed accidental leftover code
* removed all of accidental code
* added rsqrt test
* removed pow from mlops again
it was added back when resolving merge conflicts
---------
Co-authored-by: Jacky Lee <jla524@sfu.ca>
* fix syntax issues in imagenet_download.py
* use cloudpickle in cross_process to make it work in Python 3.9+
* add cross_process test
* prevent unpickling on every function call
* add cloudpickle to setup.py
* add support for args/kwargs
* new upcast works
* float4 try
* fix unaligned float4
* disallow unaligned access
* upcast dim
* maybe good now
* fix gpu half
* vstore_half4
* fix deep image bugs
* improve symbolic to fix issues
* fix symbolic
* cl test
* this maybe
* gcd of 1 is 1
* real fix for old python
* improve fuzzer
George Hotz
George Hotz
Yixiang Gao
Umut Zengin
David Hou
corranr
chenyu
nimlgen
geohotstan
Ethan Sorrell
JaSpa99
madt2709
Diogo
Steven Anderson
wozeparrot
Sieds Lykles
YiMing Han
JaSpa99
Jacky Lee
Felix
Francesco Castelli
Alex Telon
Pavol Rusnak
Karan Handa
Francis Lam
waifairer
uncommonSensor
cheeetoo
Cole Sutyak
Jacob Pradels
Alexander Edwards
Stan
Adrian Kretz
Yosef Frost
fluffy χατγιρλ
terafo
Kunwar Raj Singh
Rayan Hatout
Eli Frigo
Reza Rezvan
Daniel Hipke
Reza Rezvan